ORIGINAL ARTICLE

Annals of Nuclear Medicine Vol. 9, No. 4, 185-190, 1995

Error analysis of autoradiography method for measurement of cerebralblood flow by 123I-IMP brain SPECT: A comparison study with table

look-up method and microsphere model method

Hiroshi ITo,* Kiyoshi ISHII,** Hiroto ATsuM ^ ,** Yoshimasa INUKAI,** Shigeto ABE,** Masami SATO,**Toshifumi KINOSHITA,** Ryuta KAWASHIMA,* Shuichi ONo* and Hiroshi FUKUDA*

*Department of Nuclear Medicine and Radiology, Division of Brain Sciences,Institute of Development, Aging and Cancer, Tohoku University

**Department of Radiology, Sendai City Hospital

N-isopropyl-p-[ 123I]iodoamphetamine (IMP) has been commonly used as a cerebral blood flowtracer, but, significant clearance of IMP from the brain to the blood causes underestimation ofcerebral blood flow (CBF) as compared with true CBF when the conventional microsphere modelmethod is applied. Previously, we reported an "Autoradiography method" (ARG method) formeasuring CBF by using IMP in which this clearance effect was corrected. This method was basedon a two-compartment model (influx: K 1 , efflux: k2, K 1 /k2 = distribution volume of IMP (V,,)), theK, (corresponding to CBF) being obtained from the table which showed a correlation between CBFand the brain counts of SPECT scan with a constant V 4 value. Arterial input data used were obtainedby one point blood sampling 10 min after IMP infusion against the standard input function. In thepresent study, the ARG method was compared with the table look-up method (TLU method) andthe conventional microsphere model method (MS method) for 30 subjects. When the V,, value inthe ARG method was assumed to be 50 ml/ml, CBF values obtained by the ARG method werecorrelated well with those obtained by the TLU method (Y= 1.04X— 2.5; X: TLU, Y: ARG,r = 0.97) and those obtained by the MS method (Y = 0.82X + 12.1; X: ARG, Y: MS, r = 0.84). But,when the V,, value was assumed to be more o: less than 50 ml/ml, ARG method CBF were under-or overestimated compared with the TLU method. This indicated that the ARG method could be areliable method for CBF measurement if the V ° was determined properly. CBF values obtained bythe MS method were actually 13.2% higher than those obtained by the ARG method againstprevious studies. As reasons for this, errors in the effects of gray-white matter mixture in the ARGmethod and in estimation of the SPECT brain counts at 8 min in the MS method were considered.

Key words: IMP, SPECT, cerebral blood flow, autoradiography method

INTRODUCTION

IODINE-123 labeled N-isopropyl-p-iodoamphetamine(IMP) has been used as a cerebral blood flow (CBF) tracerfor single photon emission computed tomography(SPECT) due to its large extraction fraction and highaffinity for the brain. 1,2 But its significant clearance from

Received March 3, 1995, revision accepted May 25, 1995.For reprint contact: Ryuta Kawashima, M.D., Department of

Nuclear Medicine and Radiology, Division of Brain Sciences,Institute of Development, Aging and Cancer, Tohoku Univer-sity, 4-1 Seiryo-machi, Aoba-ku, Sendai 980, JAPAN.

the brain caused a change in IMP distribution''' andunderestimation of CBF when a conventional microspheremodel method s was applied to prolonged data acquisi-tion.^y Previously, we reported an "Autoradiographymethod" (ARG method) with IMP, a new simple ap-proach to the measurement of CBF, in which this clear-ance effect was corrected. 1°' This approach was based ona two-compartment model (influx: K 1 , efflux: k2, K,/k2 = distribution volume of IMP (V 4)) in which the V d

value was assumed to be constant, and the K, value(corresponding to CBF) was obtained from the tablewhich showed a correlation between CBF and the braincounts of the SPECT scan (mid-scan time: 40 min).

Vol. 9, No. 4, 1995 Original Article 185

uous arterial blood sampling

One point arterial blood sampling10 min

Blood sampling f//I/A- time

Scan data U IllIllIllIttIJI IIIIIIIIIIIUII timePlanar image SPECT scan SPECT scan

8 min 40 min 180 min

IMP infusion

Fig.1 The experimental protocol design shows the time schedule of scan and blood sampling. The onepoint arterial blood sampling from the brachial artery is performed at 10 min after IMP infusion fordetermination of arterial input function for TLU and ARG method. Continuous arterial blood samplingfrom the brachial artery is also performed over the 8 min after IMP infusion for MS method. Two SPECTscans are performed at mid-scan time of 40 and 180 min after IMP infusion for TLU (40 min, 180 min),ARC (40 min) and MS (40 min) methods. Brain planar images over 50 sec are obtained at 8 min afterIMP infusion for MS method.

Arterial input function was obtained by calibrating againstthe standard input function from one point arterial bloodsampling at 10 min after intravenous infusion of IMP. Ithas been considered that the CBF value obtained by theARG method would vary according to the assumed V d

value, and the Vd value would be dependent on theintrinsic performance of the SPECT scanner. 1 ' Previ-ously, we also reported a "Table look-up method" (TLUmethod) with IMP. 12- ' 4 This TLU method was also basedon a two-compartment model, in which CBF and V,, paircould be obtained from two SPECT scans and one pointarterial blood sampling. The purpose of the present studywas to compare the ARG method with the TLU methodand with the conventional microsphere model method(MS method) t5 for a major patient series.

MATERIALS AND METHODS

SubjectsSPECT studies were performed in 30 subjects including19 patients with cerebral contusion, 3 with cerebrovascu-lar disease, 2 with hypoxic brain, 2 with carbon monoxidetoxicosis and 4 normal volunteers. None of the patientshad any heart or lung disease. A informed consent wasobtained from each subject.

SPECT studyTwo SPECT scans were performed, at 40 min and 180 minof mid-scan time, after intravenous infusion of 222 MBqIMP for 1 min. Planar brain images for 50 sec wereobtained at 8 min after IMP infusion for the conventionalmicrosphere model method (MS method) (Fig. 1). TheSPECT scanner used was a Neurocam (Yokogawa Medi-cal Systems Corp., Tokyo, Japan),' 6 equipped with athree-head rotating gamma camera. In-plane resolutionwas 9 mm full width at half maximum (FWHM), and

axial resolution was 10 mm FWHM with low energy highresolution (LEHR) collimators. The SPECT scan proto-col acquired 64 projections at 50 sec per projection with120° rotation of the camera. Reconstruction was per-formed by filtered backprojection with a Butterworthfilter (cutoff frequency 0.45 cycle/cm, power factor 10).'^Attenuation correction was made numerically by as-suming the object shape to be circular or elliptical andthe attenuation coefficient to be uniform (0.12 cm')(Sorenson's method'"). The scattered photons were notcorrected. Image slices were set up parallel to the orbito-meatal (OM) line and obtained at 8 mm intervals throughthe whole brain.

One point arterial blood sampling from the brachialartery was performed at 10 min after IMP infusion. Radio-activity of the whole blood was measured with a wellcounter and was used for calibration against the standardinput function to provide an arterial input function for theautoradiography method (ARG method) and the tablelook-up method (TLU method). Continuous arterial bloodsampling at a constant rate from the brachial artery wasalso performed during the first 8 min after IMP infusion,and octanol extracted radioactivity was measured for theMS method (Fig. 1).

A cross calibration scan was performed with an ellipticcylindrical uniform phantom (long axis: 19 cm, short axis:14 cm inner diameter) for calibrating the relative sensi-tivities of the SPECT scanner and the well counter system.

Image analysisRegions-of-interest were placed in the cerebellum, pons,thalamus, putamen, centrum semiovale and cerebral cor-tex including frontal, temporal, parietal and occipitallobes on the 40 and 180 min SPECT images. The shape ofregions-of-interest was circular: 35 mm diameter for thecerebellum, and elliptic with short axes of 16-25 mm and

186 Hiroshi Ito, Kiyoshi Ishii, Hiroto Atsumi, et al Annals of Nuclear Medicine

long axes of 25-50 mm for other regions. were also calculated by microsphere model analysis asfollows:

TheoryAutoradiography method (ARG method) 1°": In thismethod, a two-compartment model was employed in linewith previous reports. 6-8,19,20

dCb(t) = K, •C0(t) — k2'Cb(t) (1)

dt

where

Ch(t): concentration of radioactivity in the brainC a(t): arterial input functionK,: influx rate constant (ml/ml/min)k2 : efflux rate constant (1/min)

In this study, we assumed the first-pass extraction fractionof IMP to be equal to 1 1,2 .21 and therefore, K, equals CBF.The ratio of K, to k2 is called the distribution volume ofIMP (V,, (ml/ml)).

Solving Eq. 1 provides:

Cb(t) = Ki-Ca(t) ©e k2 ' (2)

where © denotes the convolution integral.For a given V d value (= K 1 /k2) and a given arterial input

function, C„(t), Eq. 2 provides a unique relation betweenCh(t) and K,; then K, values (corresponding to CBF) areobtained. The arterial input function, C a(t) is obtained'bycalibration against the standard input function by usingthe arterial blood radioactivity gained from the one pointsampling.

Table look-up method(TLUmethod) 12- ' 4: In this method,a two-compartment model was also employed. For thismethod, two SPECT scans were performed. The modelequation (Eq. 2) can therefore be expressed for each scan.

Cb(t) = K,'Ca(t) ®e k2 `° (3a)

Cb(td) = K,'Ca(td) ©e kz'd (3b)

where to and td are mid-scan times at first and second scans,respectively. Calculating the ratio of Eq. 3a to Eq. 3bgives:

Cb(tc) — Ca(tc) ©e k2 tc

Cb(td) Ca(td) ® e k2 d(4)

For a given arterial input function, C^(t), the radioactiv-ity ratio of the first to second scans (the right side of Eq.4) can be considered to tabulate as a function of k 2 . Then,for a given radioactivity ratio of first to second scans, thetable look-up procedure provides a corresponding k 2

value, and the K, value can be calculated by inserting thisk2value into Eq. 3a. The arterial input function is obtainedby one point arterial blood sampling as in the ARGmethod.

Microsphere model method (MS method) 15: CBF values

Cb Cb-Rf = ammo = (5)

Ca

Ca(t)dtf0min

where

f: CBF (ml/ml/min)C,,: concentration of radioactivity in the brain at 8

min after IMP infusionCa(t): arterial input functionR: constant arterial blood sampling rate (ml/min)Ca: total octanol extracted radioactivity of the blood

withdrawn over 8 min

In this study, SPECT data at 8 min (C,,) were 40 minSPECT data calibrated by the count ratio of 8 min/40 minwhole brain planar images.

Simulation of the effects of gray-white matter mixtureThe limited spatial resolution of the SPECT scannercauses a gray-white matter mixture in regions-of-interest.The effects of gray-white matter mixing on CBF valuescalculated by the ARG method were evaluated. 22 Braincounts in heterogeneous tissue on a 40 min SPECT scanwere generated as mixtures of gray and white matter. CBFvalues for the gray and white matter were assumed to be80 and 20 ml/100 ml/min, respectively. The Vd value forgray and white matter was assumed to be the same as 40,50 or 60 ml/ml. The difference between true CBF values(= 80 ml/100 ml/min x gray matter fraction + 20 ml/100ml/min x white matter fraction) and CBF values calcu-lated by the ARG method by using the generated hetero-geneous tissue radioactivity were estimated where thefraction of gray matter per given region-of-interest variedfrom 0 to 100%. In this simulation, the arterial inputfunction was the standard input function used for the ARGmethod.

RESULTS

Figure 2a-2c show correlations between CBF values byTLU method and those by the ARG method in which theV, values were assumed to be 40, 50 and 60, respectively.When the V d value in the ARG method was assumed to be50 ml/m1, the best correlation and good linearity betweenthe TLU and ARG methods were observed (Y = 1.04X —2.5; X: TLU, Y: ARG, r = 0.97) (Fig. 2b) and these valuesfor the two methods were consistent. However, when theVd value was assumed to be 40 or 60 ml/ml, the ARGmethod CBF was over- (Fig. 2a) and under- (Fig. 2c)estimated compared with the TLU method CBF, respec-tively, in particular in the hyperperfusion region. Themean CBF values obtained by the ARG method for allregions-of-interest data were 39.1 ± 10.3, 36.3 ± 8.6 and

Vol. 9, No. 4, 1995 Original Article 187

^ 100

EE ^ 80

II £ 60

> OO

" t- 40LL i-m E6 20

Cra o

• iII

¤¤

¤

LL 80mU

C 60d.O—

p 40

d

H 200V

0

Y=0.82X+12.1r =0.84

•

0 20 40 60 80 100

TLU-CBF (ml/100m1/min)a

Y=1.04X-2.5r=0.97

U.

U0

0̀

o -5

-10

Mm0 20 40 60 80 100

TLU-CBF (mi/100ml/min)

b

»._..»»»» Vd=40

--- Vd=50

----- Vd=60

0 20 40 60 80

ARG-CBF (ml/100mi/min)

Fig. 3 Correlation between CBF values evaluated by ARGmethod in which the Vd value is assumed to be 50 ml/m/ andthose from MS method. A good correlation is obtained(Y = 0.82X + 12.1; X: ARG, Y: MS, r = 0.84). However, overes-timation of CBF values is observed with MS method as com-pared with those from ARG method. The dot line is identity.

10

Gray-CBF = 80 ml/100ml/minWhite-CBF = 20 ml/100mI/min

-15

100EE ^_ 80

Eii= 60

Or 40

fn E20

C,Ma 0

a

^ 100

EE 80

o=so60

a E>0•--. 40LLm £

20

or.0 20 40 60 80 100

TLU-CBF (ml/100mI/min)C

Fig.2 Figures show correlations between CBF values by TLUmethod and those by ARG method in which the V,, value isassumed to be 40 (a), 50 (b) or 60 (c). When the V d value in ARGmethod is assumed to be 50 ml/ml, the best correlation isobserved (Y= 1.04X — 2.5; X: TLU, Y. ARG, r = 0.97) (b). Thedot line is identity.

34.8 ± 7.8 ml/100 ml/min (± S.D.) when the V d valueswere assumed to be 40,50 and 60 ml/ml, respectively. Themean CBF value by the TLU method was 37.4 ± 8.1 ml/

0 20 40 60 80 100

Gray matter fraction (%)

Fig. 4 Simulation of the effects of gray-white matter mix-ing on CBF values calculated by ARG method. For each V d

value, CBF values calculated are systematically underestimated(-7.6% for a gray matter fraction of 50% and a Vd of 50 ml/ml).

100 ml/min (± S.D.).The mean V,, value evaluated by the TLU method in X-

ray CT normal density regions was 48.7 ± 9.15 ml/ml(± S.D.). There was no significant difference betweengray and white matter in V,, values.

A good correlation was obtained between CBF valuesevaluated by the ARG method in which the. V d value wasassumed to be 50 ml/ml and those by the MS method(Y= 0.82X + 12.1; X: ARG, Y: MS, r=0.84) (Fig. 3).However, overestimation of CBF values was observedwith the MS method as compared with the ARG method.Mean CBF values evaluated by the ARG method were13.2% lower than those by the MS method (mean MSmethod CBF: 41.8 ± 8.5 ml/100 ml/min (± S.D.)).

Figure 4 shows the effects of gray-white matter mixingfor CBF values calculated by the ARG method. For eachVd value, CBF values calculated were systematically

188 Hiroshi Ito, Kiyoshi Ishii, Hiroto Atsumi, et al Annals of Nuclear Medicine

underestimated when the fraction of gray matter wasvaried from 0 to 100% (-7.6% for a gray matter fractionof 50% and a V 4 of 50 ml/ml).

DISCUSSION

The best correlation between the TLU and ARG methodswas observed (Fig. 2b) when the V4 value in the ARGmethod was assumed to be 50 ml/ml which was almost thesame as the mean V4 value evaluated by the TLU method.These ARG method values also closely agreed with theMS method values (Fig. 3). However, this V ° value waslarger than those in previous studies, i.e., 22 in man, 5 22in rat'-; and 29.6 in man' (ml/ml). As reasons for thisdiscrepancy, the differences in the reconstruction algo-rithm, the attenuation correction method and the scatteredphotons correction method for each SPECT scanner sys-tem were considered)° These indicated that the V 4 valuein the ARG method should be determined for each SPECTscanner system by using an other method, e.g., the TLUmethod, since the reconstruction algorithm, the attenua-tion correction method and the scattered photons correc-tion method were different for each SPECT scannersystem.

V4 values in the normal brain tissue might be almostconstant, but it was reported that V d values in the patho-logical regions, e.g., cerebral infarction were lower thanthose in normal regions. 11,11 . 2° In the present study, allregions-of-interest in the pathological regions, e.g., cere-bral contusion and cerebral infarction showed low V,,values and hypoperfusion. Since in hypoperfusion re-gions accurate CBF values could be calculated by theARG method in any assumed V d values, 10 a good correla-tion was observed between the ARG and TLU methods(Fig. 2b). However, if there were hyperperfusion lesionswith low or high V4 values, the error in measuring CBF bythe ARG method would be significant.

While the MS method was routinely used as a methodfor measuring CBF by means of IMP, the underestimationof CBF was caused due to significant clearance of IMPfrom the brain." ,24 In this study, a good correlation wasobtained between CBF values evaluated by the ARGmethod and those by the MS method (Fig. 3), suggestingequivalent applicability. However, CBF values obtainedwith the MS method were actually 13.2% higher thanthose obtained with the ARG method against previousstudies." ,24

One possibility is error in the effects of gray-whitematter mixture. Simulation of the effects of gray-whitematter mixture indicated differences between true CBFvalues and CBF values calculated by the ARG method. Inthis simulation, CBF values calculated were systemati-cally underestimated (-7.6% for a gray matter fraction of50% and a V ° of 50 ml/ml) (Fig. 4). On the other hand,there were no effects of gray-white matter mixing on CBFvalues calculated by the MS method, because the correla-

Lion between the brain radioactivity and CBF value islinear in the microsphere model (Eq. 5). 22

Another potential source of error is in estimation of theSPECT brain counts at 8 min in the MS method with countratios of 8 min/40 min whole brain planar images and 40min SPECT data. 2¢26 This error would be caused by non-linearity of 40 min SPECT counts due to a significantclearance of IMP which had been described in our re-port. 24 An additional potential source of error is thedetermination of the arterial input function, i.e., correc-tions for time delay and dispersion of input for twomethods27-21 and unknown errors due to differences in thearterial input curve shape for each subject for the ARGmethod.24

In conclusion, the best correlation between the TLUand ARG methods was observed when the V 4 value in theARG method was assumed to be 50 ml/ml which wasalmost the same as the mean V d value evaluated by theTLU method. These ARG method values closely agreedwith conventional MS method values, suggesting equiva-lent applicability. This indicates that the V d value in theARG method should be determined properly for eachSPECT scanner system by using an other method, e.g., theTLU method. CBF values obtained by the MS methodwere actually 13.2% higher than those obtained by theARG method against previous studies. As reasons for this,errors in the effects of gray-white matter mixture in theARG method and in estimation of the SPECT brain countsat 8 min in the MS method were considered.

ACKNOWLEDGMENTS

We are greatly indebted to the staff of Sendai City Hospital andthe Institute of Development, Aging and Cancer, Tohoku Uni-versity.

This study was supported by a Grant-in-Aid for ScientificResearch (05454297) from the Japanese Ministry of Education,Science and Culture.

REFERENCES

1. Winchell HS, Baldwin RM, Lin TH. Development of 1- 123-labeled amines for brain studies: localization of 1-123iodophenylalkyl amines in rat brain. J Nucl Med 21: 940-946, 1980.

2. Winchell HS, Horst WD, Braun L, Oldendorf WH, HattnerR, Parker H. N-isopropyl-[' 231] p-iodoamphetamine: single-pass brain uptake and washout; binding to brain synapto-somes; and localization in dog and monkey brain. J NuclMed 21: 947-952, 1980.

3. Creutzig H, Schober 0, Gielow P, Friedrich R, Becker H,Dietz H, et al. Cerebral dynamics of N-isopropyl-(" 3I)p-iodoamphetamine. JNucl Med 27: 178-183, 1986.

4. Nishizawa S, Tanada S. Yonekura Y, Fujita T, Mukai T,Saji H, et al. Regional dynamics of N-isopropyl- ( 123 I)p-iodoamphetamine in human brain. JNucl Med 30:150-156,1989.

5. Kuhl DE, Barrio JR, Huang SC, Selin C, Ackermann RF,

Vol. 9, No. 4, 1995 Original Article 189

Lear JL, et al. Quantifying local cerebral blood flow by N-isopropyl-p- [ 123 ljiodoamphetamine (IMP) tomography. JNucl Med 23: 196-203, 1982.

6. Greenberg JH, Kushner M, Rango M, Alavi A, Reivich M.Validation studies of iodine-]23-iodoamphetamine as acerebral blood flow tracer using emission tomography. JNucl Med 31: 1364-1369, 1990.

7. Murase K, Tanada S, Mogami H, Kawamura M, MiyagawaM, Yamada M, et al. Validation of microsphere model incerebral blood flow measurement using N-isopropyl-p-(123I) iodoamphetamine. Med Phys 17: 79-83, 1990.

8. Yokoi T, Iida H, Itoh H, Kanno 1. A new graphic plotanalysis for cerebral blood flow and partition coefficientwith iodine-123-lodoamphetamine and dynamic SPECTvalidation studies using oxygen-l5-water and PET. J NuclMed 34: 498-505, 1993.

9. Yonekura Y, Nishizawa S, Mukai T, Iwasaki Y, FukuyamaH, Ishikawa M, et al. Functional mapping of flow and back-diffusion rate of N-isopropyl-p-iodoamphetamine in hu-man brain. J Nucl Med 34: 839-844, 1993.

10. Iida H, Itoh H, Nakazawa M, Hatazawa J, Nishimura H,Onishi Y, et al. Quantitative mapping of regional cerebralblood flow using Iodine-123-IMP and SPECT. J Nucl Med35: 2019-2030, I994.

11. Iida H, Itoh I-I, Nakazawa M, Nishimura H, Onishi Y,Uemura K. Validation of quantitative mapping of rCBFusing I-123 IMP from a single SPECT scan with a standardinput function. J Nucl Med 35: 191P, 1994.

12. lida H, Itoh H, Bloomfield PM, Munaka M, Higano S.Murakami M, et al. A method to quantitate cerebral bloodflow using a rotating gamma camera and iodine-123iodoamphetamine with one blood sampling. Eur J NuclMed 21: 1072-1084, 1994.

13. Iida H, Itoh H, Munaka M, Murakami M, Higano S, UemuraK: A clinical method to quantitate CBF using a rotatinggamma camera and I-123-amphetamine (IMP) with oneblood sampling. KAKU IGAKU (J Nucl Med) 33: 963P,1992.

14. Itoh H, Iida H, Murakami M, Bloomfield PM, Miura S,Okudera T, et al. A method for measurement of regionalcerebral blood flow using N-isopropyl-p- [ 123l]iodoamphet-amine ( 1231-IMP) SPECT; two scans with one point bloodsampling technique. Jpn J Nucl Med 29: 1193-1200, 1992.

15. Matsuda H, Seki H, Sumiya H, Tsuji S. Tonami N, HisadaK, et al. Quantitative cerebral blood flow measurementsusing N-isopropyl-(iodine I23)p-iodoamphetamine andsingle photon emission computed tomography with rotatinggamma camera. Am J Physiol Imag 1: 186-194, 1986.

16. Kouris K, Jarritt PH, Costa DC, Ell PJ. Physical assessmentof the GE/CGR Neurocam and comparison with a singlerotating gamma-camera. Eur J Nucl Med 19: 236-242,1992.

17. Budinger TF, Gullberg GT, Huesman RH. Image recon-struction from projections. Herman GT (ed.), New York,Springer-Verlag, p. 197, 1979.

18. Sorenson JA. Quantitative measurement of radioactivity invivo by whole-body counting. In Instrumentation in Nuclear

Medicine, Hine GJ, Sorenson JA (eds.), New York, Aca-demic Press, pp. 311-348, 1974.

19. Itoh H, lida H, Bloomfield PM, Murakami M, Higano S,Munaka M, et al. A technique for rapid imaging of regionalCBF and partition coefficient using dynamic SPECT and 1-123-amphetamine (IMP). J Nucl Med 33: P911, 1992.

20. Ito H, Iida H, Bloomfield PM, Murakami M, Inugami A,Kanno 1, et al. Rapid calculation of regional CBF anddistribution volume using N-isopropyl-p- [ 123lJiodloamphet-amine ( 123 I-IMP) and dynamic SPECT. JNucl Med 36: 531-536, 1995.

21. Murase K, Tanada S, Inoue T, Ochi K, Fujita H, Sakaki S,et al. Measurement of the blood-brain barrier permeabilityof 1-123 IMP, Tc-99m HMPAO and Tc-99m ECD in thehuman brain using compartment model analysis and dy-namic SPECT. JNucl Med 32: P911, 1991.

22. Huang SC, Mahoney DK, Phelps ME. Quantitation inpositron emission tomography: 8. Effects of nonlinear pa-rameter estimation on functional images. J Comput AssistTomogr 11: 314-325, 1987.

23. Lear JL, Ackermann RF, Kameyama M, Kuhl DE. Evalu-ation of [ 123I]isopropyliodoamphetamine as a tracer forlocal cerebral blood flow using direct autoradiographiccomparison. J Cereb Blood Flow Metab 2: 179-185, 1982.

24. Ito H, Ishii K, Atsumi H, Kinoshita T, Kawashima R, OnoS, eta!. Error analysis of Table look-up method for cerebralblood flow measurement by ' 231-IMP brain SPECT: Com-parison with conventional microsphere model method. Ann

Nucl Med 9: 75-80, 1995.25. Takahashi N, Ohkubo M, Odano I, Sakai K. A problem of

quantitative measurement of regional cerebral blood flowusing microsphere model and N-isopropyl-p-[' 23 1]iodoam-phetamine (IMP): Comparison with ' 33Xe SPECT and se-quential dynamic ' 23 l-IMP SPECT. KAKU IGAKU (Jpn JNucl Med) 31: 319-326, 1994.

26. Takahashi N. Quantifying regional cerebral blood flow withN-isopropyl-p[ 12'I]iodoamphetamine by ring-type single-photon emission computed tomography: Validity of a methodto estimate early reference value by means of regional braintime-activity curve. Ann Nucl Med 8: 253-258, 1994.

27. lida H, Kanno 1, Miura S, Murakami M, Takahashi K,Uemura K. Error analysis of a quantitative cerebral bloodflow measurement using H 2 15O autoradiography and positronemission tomography, with respect to the dispersion of theinput function. J Cereb Blood Flow Metab 6: 536-545,1986.

28. lida H, Higano S. Tomura N, Shishido F, Kanno I, Miura S.et al. Evaluation of regional differences of tracer appear-ance time in cerebral tissues using [ 15O]water and dynamicpositron emission tomography. J Cereb Blood Flow Metab

8: 285-288, 1988.29. Iida H, Kanno I, Miura S, Murakami M, Takahashi K,

Uemura K. A determination of the regional brain/blood.partition coefficient of water using dynamic positron emis-sion tomography. J Cereb Blood Flow Metab 9: 874-885,

1989.

190 Hiroshi Ito, Kiyoshi Ishii, Hiroto Atsumi, et al Annals of Nuclear Medicine